Professional Documents
Culture Documents
Harvard Astrophysics Book
Harvard Astrophysics Book
Introduction
b Systems are often complicated and we cannot always conduct exper iments to isolate the relevant variables parameters. The observations that astronomy provides are often very inaccurate and incomplete so we often need only work to low accuracy order of magnitude calcula tions. A factor 2 error in the theory is excellent in some areas a factor 102 is good in others. Sometimes the data are embarrassingly precise and we have to do much better e.g. in solar system studies. c Astrophysics covers a broad spread of physics including topics that normally do not appear in a physics curriculum e.g. uid mechanics magnetohydrodynamics radiative transfer etc. Many phenomena that cannot be reproduced in the laboratory occur in astrophysics e.g. ux freezing gravitational collapse so astrophysical interpretations may o er the only paths to some parts of physics. d But note we do not allow new physics to be involved very often. As trophysics corresponds to an application of the standard laws of physics to the Universe as a whole. We assume the universality literally of the laws of physics in order to make any progress at all and to be allowed to call the subject astrophysics The extent to which our laboratory laws extend to the distant objects in the Universe and pro vide explanations of phenomena there provides evidence for this point of view. e Astrophysics is intimately connected with what astronomy is able to ob serve. Until recently there was little high energy astrophysics because there was no high energy x ray and ray astronomy. Now there is and we can see much hotter parts of objects than we knew about before. 2
f Astronomy and astrophysics generally assumes the Copernican princi ple WE ARE IN NO FAVORED POSITION IN THE UNIVERSE. so the Universe near us and physics near us are typical of contemporane ous parts of the Universe unless of course we nd factual evidence to the contrary and this should be in each such case only by rare good or bad luck. When does astronomy turn into astrophysics No hard and fast line but where physical laws not just geometric arguments or logic are applied.
Earth
Superior conjunction
Quadrature East
East
If one assumes circular orbits and constant velocities along orbits can check this observationally then the dates of opposition quadrature greatest elongation conjunction give the relative geometry. 3
Cayenne
Not to scale
Mars
Mars
They got the answer wrong or right to 10. Technology aside Micrometer sights were rst used in the 1670s. Halley in 1716 age 60 pointed out that the transit of Venus in 1761 and 1769 could be used. He knew he would be dead by then and told all young astronomers about it. These transits are quite infrequent next ones were 1874 1882 June 7 2004 June 5 2012.
0.72 AU Earth d
D
Venus
Sun
1 AU
D d
0 72 0 28
How to measure small separation from opposite sides of globe Not by subtracting measurements of height from top of sun too imprecise. Rather take the average height to get the geometry and use the di erence of transit ingress egress times or durations to infer the small di erences. Order of magnitude ignoring magni cation factors geometry etc.
GM m R2
Sun
So
Candle
How to compare to a candle Not easy before electrical measurements In modern units solar illuminance is 127000 lux. 6
1 lumen 1 lux
1 cm2 blackbody light emitted by 60 at the melting point of platinum 2044 K 1 lumen m2 1 candle at 1 m
A bright mV 0 star is 2 5 10 6 lux. A good VCR camera with a charge coupled device CCD works down to 5 lux. Use a pinhole to get into candle comparison range
L pinhole image of sun. Area A'
area A
A0
D2 4
L 1 1 rad 4 2 57 3 L2 6 0 10 5 L2 6 0 10 A
2
A0 A
Taking L 20 m 2 103 cm A 1 cm2 10 2 cm2. This gives A0 A 24000 which gets sun down to 5 candles at 1 m or 1 candle at 45 cm. So by this and similar techniques around time of Huygens 1650 people knew the candlepower of the Sun
2 127000 candles 11AU m
2
candles
We now know that the actual distance to Sirius is 550000 AU 8.8 l.y.. It is much brighter intrinsically than the sun and Huygens calibration of it in candles was also not accurate.
Huygens also thought that Sirius modern mV 1 45 was as bright as a candle at 80 m. So if it was same intrinsic brightness as Sun s DSirius 127000 29000 AU DSun 1 802 1 2 light year.
Sun
close star If this angle is _ 1 arc sec the star is at a distance of 1 _ parsec [parallax second]
Since 1 arcsec
So 1 pc 206265 AU 3 26 ly. With modest astrometric telescopes and lots of e ort direct parallaxes are possible out to 10 50 pc 0.1 to 0.02 arcsec. Beyond that indirect methods must be used. 8
First parallaxes date from 1839 1840 Bessel von Struve. Technology note both used lenses made by Fraunhofer Bavaria.
1 m 2 pan of water with black bottom and known depth of water Measure rate of temp. increase when Sun is overhead.
Correcting for absorption in atmosphere1 the answer the solar constant is about 1400 W m2. So
1 AU Sun
L 41 AU2 L
How Try di erent times of day and measure secant e ect e.g.
We can usually measure small angles 1 much better than large angles say 45 . We nd accuracies of Optical Radio 0.00001 0.00000001 best best Very Long Baseline Interferometry VLBI limited by spacecraft limited by properties of ray detectors
X ray ray
200 1
1 radian 57 3
or RA DEC
The Earth is a pretty good gyroscope so its axis points a constant direction in inertial space the North Celestial Pole NCP. Declination or DEC is measured from 90 at that pole to 90 at the South Celestial Pole. The longitude like coordinate is called right ascension or RA and is measured confusingly in hours 24h 360 . The zero point is at the vernal equinox where the Sun is in the sky at the beginning of Spring also called the rst point of Ares. RA increases in number in the direction the sky moves as the Earth turns. That is a xed telescope sees increasing RA positions with time 1 hour RA per hour of sidereal time. One peculiarity is that objects at rest on sky have RA Dec which vary very gradually with time because of the Earths precession. This occurs because the spin axis of the earth is not aligned with the Earth Sun orbital plane.
23.5 23.5
"now"
11
spi
nv ect
or
If the earth was a sphere this would not a ect the Earths spin axis there would be no coupling but tides and rotation distort the earth so it feels a net torque from the gravity of the sun moon. This torque makes the polar axis direction to NCP precess with a period of 26 000 years.
23.5
Every 26 000 years the RA goes through a big loop and DEC changes by 23 5 . Positions of objects in the sky change by tens of arcseconds per year. This is easily detectable as we can measure the angle to 10 3 arcsec.
Coordinates quoted for objects are therefore refered to a particular date or standard epoch to remove e ects of precession. Standards are B1950 going out of use J2000 coming in Here the B and J refer to technicalities of the model while 1950 and 2000 are the reference dates typically noon on January 1 of the date.
Galactic Coordinates l b
The equator of the galactic coordinate system is the galactic plane the Milky Way. l the longitude coordinate is zero in the direction of the Galactic Center. b the galactic latitude goes from 90 at the North Galactic Pole NGP down to 90 at the SGP. 12
is ok when is not small but for small separations it becomes inaccurate angles pile up around cos 1. In that case use the construction
^ r1 (1/2)
(1/2) |^ r1 - ^ r2|
which yields
tan 1 2
jr jr
1 1
rj rj
2 2
analogy with the de nition of radians the unit of solid angle the steradian measures the area on the unit sphere that the object covers. Since a sphere has area 4r there are 4 steradians in the whole sky. The corresponding number of square degrees is
2
1 cm 2
distance D
You can see that if the source emits isotropically same in all directions the relation between ux F and total luminosity L erg s is L F 4D
1 2
14
Ideally Flux would be the total energy emitted but in practice we measure di erent wavelength bands with di erent instruments or detectors. In such cases we label it by a letter telling which band e.g. V visual and it then means energy per area per time arriving in that band. The most common bands are U ultraviolet B blue V visual R red I J K infrared.
100 Transparency (%) 80 60 40 20 0 3000 4000 5000 6000 A Wavelength
Light Transmission Through UBV Filters. This graph shows the wavelength ranges
Band U B V R I J K Central Wavelength 3650 A 4400 A 5500 A 7000 A 9000 A 1.25 2.2 E ective width de ned 680 2400 A 980 A 890 A 2200A A 0.38 48 by area under curve log f erg cm 2 s 1 1 4.37 4.18 4.42 4.76 5.08 5.48 6.40 for m 0
over which the standardized U B and V lters are transparent to light. The U lter is transparent to the near ultraviolet. The B lter is transparent from about 380 to 550 nm and the V lter is transparent from about 500 to 650 nm.
Note that 1000 A 100 nm 0.1. We write F F F etc. for in band uxes. When we need a word for real total energy we say bolometric so
U B V
bol
To interpret the last line in the above table we need to know about the magnitude scale. 15
2 5 log F
2 5 log F constant
1
Once we decide on magnitude for one star all others are determined. Since stars have di erent colors well learn why later we must compare them in a speci c color band i.e. look at them through lters as shown in 2.2.1. The naked eye is esentially a V lter. Thus like uxes magnitudes are written with a subscript indicating band m m etc. Sometimes this is written just V B etc. The zero point constant derives historically from the ancient Greeks who named some bright stars as being of the rst magnitude what we would now call V 0
V B
Arcturus Boo V 0 06 B V 1 23 Vega Lyr V 0 04 B V 0 0 Capella Aur V 0 8 B V 0 79 Note that larger magnitude means dimmer Betelgeuse Ori V 0 8 B V 1 85 Aldebaran Tau V 0 85 B V 1 53 16
Larger color di erence means redder color. With good eyes and a dark sky not Cambridge you can see stars down to V 6. How much energy do we receive from Betelgeuse in the V band From the previous table for V 0
10
4 42
3 4 10 10
6
0 40 8
erg 1 6 10 cm s
6 2
You can see that the general relation for any band X e.g. X
UBV
FX 10
CX W 10 X
04
mX
where CX is the bands log f for mX 0 WX is its e ective width and mX is the apparant magnitude of the object in question. Incidentally since 10 0 398 is pretty close to e 0 367 magni tudes are not too di erent from e folds. This is useful for calculating small magnitude di erences in your head e.g.
04 1
0 03 mag e
0 03
1 0 03 1 03
So a di erence of 0.03 mag is about a 3 ux di erence and so forth. Another useful fact if you like decibels is that 1 mag 4 dB. Therefore e.g. 2.5 mag 10 dB factor of 10 in intensity.
17
m M 5 log 10Dpc M 5 log D 5 the 2.5 on log ux becomes a 5 on log distance. The quantity m M 5 log D 5
is called the distance modulus. You should now be able to derive the relation between absolute magnitude M and physical luminosity L in a given color band for an object. These are properties of the object not of its distance. If we know L or M for an object then measurement of m gives D the distance. A standard candle is a hoped for class of objects which has a luminosity absolute mag which can be determined easily without knowing its distance. Notice that color di erences e.g. B V are independent of distance and are equal to M M e.g. Here are the masses absolute magnitudes and so called spectral types of stars on the main sequence well learn more about this later. Spectral type O B A F G K M Typical Mass units of M 40 6 2 1.5 1.0 0.7 0.3 M 5 8 1 1 2.0 3.4 5.1 7.3 11.8 The Sun is a G star at 10 pc it would be barely visible m 5 1.
B V V V
distance
2.2.4 Spectra
Atoms in a gas can emit light at speci ed frequencies called spectral lines. They can also absorb light at these frequencies. Whether they on balance emit or absorb depends on their temperature ionization density etc. So a spectrum might look like 18
emission lines
f
"Continuum level" absorption lines 0
Wavelength
The units of f are ergs per cm collecting area per s per wavelength. This is the same f that we previously saw in the UVB magnitude table.
2
vr
vt velocity v
o z vr in nonrelativistic limit o o c Between successive emissions of crests time intervals c the emit ter moves a distance v c while the wave moves a distance so v c v c
0 0 0 0 0 0 0
19
0 v c
position now
Lab Observed
blue 0 red
ii proper motion Here we observe motion of object in plane of the sky over time sometimes a very long time. We get vt v sin from the rate of change of position on the sky if we know the distance to the object. We generally do know the distance from parallax measurements q.v. because we can only see proper motions for nearby stars. E.g. to see appreciable motion say 100 century given that the typical velocity of a star is 10 km s we can only measure vt for stars within distances 2 kpc about one fth the distance to the center of the galaxy.
1
20
We can do better with VLBI radio techniques using masers an as tronomical analog to a laser because 1 we can determine angles more accurately and 2 maser velocities are larger than veloci ties of nearby stars. This has been used to get the distance to the galactic center.
ii triangulation again okay for Venus Moon. Taken with radar it leads to measurements of the AU.
21
Angles p are small 100 and the distance D R p. If R 1 AU and p is measured in seconds of arc then D comes out in parsecs. 1 pc distance at which 1 AU subtends 1 arcsec 3 086 10 cm.
18
iv inverse square law and standard candles If you know that a certain object has luminosity L e.g. by look ing at similar objects near you L in erg sec or watts then the distance of that object is given from its observed ux
L 4D
D
Source
22
This method works well for Cepheids variable stars whose variability period obtained from long time observations is related to L. Thus we use the measured period to infer L and combine L with the observed brightness to get D. RR Lyrae stars another kind of variable star and Tully Fisher galaxies in which the galaxys luminosity is inferred from its rotational velocity are used in a similar way. Note that extinction the absorption of light by intervening material can cause problems. v angular sizes of standard rods Same idea as iv nd a distant object that is the same as a nearby one for which you know the size. If the nearby object has a true size L and the distant object has angular size then
D
vi Hubble law galaxies and beyond
Use the concept that the Universe is expanding. If the Hubble law has been calibrated for you this involves getting measurements of distances of objects by independent means then a radial velocity measurement leads to the distance from
vr HoD
where vr is the Doppler velocity and Ho 50 100 km s pc . This is usually written Ho 100 h km s pc where h 0 5 1.0. But note that peculiar velocities and gravitational redshifts might cause problems.
1 1 1 1
23
Note that the use of method vi gives an uncertainty of a factor 2 in the distance scale This is a real problem for astrophysics. Ho is not well known because of the di culty in determining distances to distant galaxies from methods iv and v.
24
Radiation
Virtually all the information we get in astronomy comes to us as electromag netic radiation radio infrared visible UV X ray or gamma ray in other words as photons.
E pc
Note that p is the magnitude of the 3 vector momentum p
p jpj
each point that matters but also the distribution of which way they are going and of what energies they have. Since the momentum vector p encapsulates both the latter concepts direction and energy it is useful to think at some instant in time of each photon being at a point in a 6 dimensional space of position and momentum x y z px py pz I can only draw 2 of the dimensions
dx
dpx
p x
The phase space density N is the number of photons dN per 6 dimensional unit phase space
px
26
1 E 2dE d c3 So for photons in a xed energy interval dE around E d3p is proportional to the solid angle element d that encompasses their directions and to E 2dE . We should here say a word about the notation d for solid angle it is measured in steradians a unit such that the full 2 sphere is 4 steradians i.e. d 4 sphere
Then
d3p p2 dp d
In spherical polar coordinates d decomposes into the two orthogonal direc tions d sin dd dcos d
d
sin d
If we consider photons moving in one direction and normally incident on a detector then d3x is related to their ux
27
Q dz
Photons in volume dx dy dz will slam into wall in a time dz c. So number per volume 1 c number per area per time or
d3 x c dx dy dt c d2 dt
Why do we care Because of a very famous subtle theorem of mechanics
normal area 2
d
z
28
Example Compare photons close to and far from an extended source like
a star
star
photons here are crowded in physical space (high density) but spread out in solid angle (direction)
photons here are diffuse in physical space (low density) but highly collimated in angle (high density).
So dilution of photon density is always accompanied by increased colli mation in a precise quantitative relationship
3.1.3 Brightness
Lets compute the total energy dE in a small volume of phase space
dE EdN
E N d3 x d3p
E N cd2 dt
1 E 2 dE d c3 E 3N c2 d2 dt dE d
Id2 dt dE d
Here we have de ned
I N E 3 c2
We see that I is the energy in photons per normally incident area per time per solid angle per photon energy interval dE. I is called the speci c intensity or brightness of a radiation eld at a particular point in a particular direction at a particular energy. 29
Note that since N is conserved Liouville and E is constant for each photon I is conserved too If you look at a source while moving farther and farther away its brightness in this context often called surface brightness doesnt change. It gets smaller but not dimmer per unit apparent size. All optical systems indeed satisfy Liouvilles Theorem. One implication is that no telescope however large can make a scene look brighter than it looks to the naked eye. It can only make it look bigger. However for very small unresolved sources like stars bigger e ectively is brighter since there is more light in the resolution element. CCD detectors and photographic emulsions dont measure surface bright ness as the eye does but rather respond to total energy deposited per area of collector. For these a bigger telescope gathering more total photons is advantageous. We will come back to this topic later.
Refraction in lenses by the principle of least time Interference e ects Finite resolution of an aperture A telescope of aperture diameter D cannot distinguish directions any ner than about
in radians D
The light from unresolved point sources is seen as blurred into a spot of about this radius.
E h
where h 6 63 10 27 erg sec is Plancks constant. One way to visualize the photon as both a particle and wave is to think of it as a propagating wave packet.
The more monochromatic small E or the photon the longer the wave packet. This is why it is easy to do interference demonstrations with laser light very monochromatic but quite di cult with white light large . The relation between length of wave packet and frequency or energy or momentum localization is of course the Heisenberg uncertainty principle. 31
Notice that
dE hd
I c2 3 h4
dE I d2 dt d d
with I having the units energy per normal incidence area per time per frequency interval per solid angle that is erg s 1 cm 2ster 1Hz 1. I is a function of position in space direction under consideration photon frequency under consideration possibly though not usually time
32
Do not be confused by these extra factors of or . They are just bookkeep ing to make per frequency and per wavelength come out consistently when describing the same amount of energy.
3.3.3 Net ux
If we want to know the total energy deposited onto a speci c element of area dA on a detector say we integrate up the speci c intensity coming from solid angle elements in all directions. However they do not all get equal weight the farther they are from normal incidence the less projected area of detector is available to them.
d
normal direction dA
33
F
I cos d
F is the energy ux per area per time per wavelength interval. In fact it is exactly the same quantity that we called f in the table of UBV photometry in 2.2.1. Note that we are always free to integrate up all the di erent frequencies and obtain total integrated ux F I
Z
In spherical coordinates d sin d d d cos d. What are the limits of integration If dA represents an opaque detector goes from 0 to 90 . On the other hand if dA is transparent then goes from 0 to 180 and the cos weighting is negative on the back side 90 180 . For an isotropic radiation eld I is not a function of or and one sees that the net ux including from the back is exactly zero. As before we can do the bookkeeping per wavelength instead of per frequency with 2 F c F c2 F
F d I d
F d I d
between and d is
u d
and the total energy density is
1 c
Z
I d d
u d
1 c
ZZ
I d d
Pressure is force per area or momentum transfer per area per time. Con sider photons bouncing back and forth between parallel perfect mirrors of area A and separation L.
Each photon contributes u E AL to the energy density since the volume is AL. Each photon gives a momentum transfer 2p each time it bounces o the mirror p goes to p so the di erence is 2p. These bounces occur on a given mirror every 2L c time so the photons contribution to the pressure is pc E u 2p P A2 L c LA LA Adding up all the photons we would nd
u but wait
We have forgotten that photons are moving at di erent angles. A photon moving at angle to the normal still contributes E AL to u but 35
Therefore
the momentum transfer is now only 2p cos and also it takes longer between bounces LC t 2cos P cos2 u
Adding up all the photons is the same as averaging an isotropic distribution over cos2
hcos2 i
RR
1
R
2 1 d 1 d 1
1 3
where we set cos . Another way to get this result is by symmetry Consider a randomly oriented unit vector x y z with x2 y2 z2 1. Since there is nothing special about the x y or z directions with respect to the unit sphere we must have
1u 3
c
r
Then
I cos d
2 0
c
0
sin cos d
where c sin 1 R r is the angle at which a ray from P is tangent to the sphere. It follows that
F
or
B 1 cos2 c B sin2 c F
2 B R r Thus the speci c intensity is constant but the solid angle subtended by the given object decreases in such a way that the inverse square law is recovered.
37
B
3.4 Telescopes
3.4.1 Astronomical Telescopes
All modern telescopes optical radio X ray are mirrors not lenses. How ever mirror optics is just lens optics folded over
Since it is easier to draw lenses we will use the left hand gure but you should mentally translate it to the right hand one. D is the aperture L the focal length. The basic rules of ray tracing are rays go through center of lens unde ected parallel rays get converged and come together at the focal length That is all we need to get
38
This is the basic prime focus astronomical telescope. The detector CCD array goes at the focus. The image scale is the ratio of physical distance at the focus to angular distance in the sky and it depends only on the focal length see gure l L or 1 L l For example the 16.8 m focal length of the Palomar 200 inch corresponds to 12 arcsec mm check this The eld of view is determined by the focal length and the size of the detector d in the gure FOV d L In the above example a CCD array of diameter 60 mm 6 cm would have a eld of view of arcsec 1 arcmin 60 mm 12 1 mm 60 arcsec 12 arcmin 0 2
What is the relation between the surface brightness of an object in the sky a planet or galaxy say and the ux energy per area per time in its image on the detector We multiply I by the area of the lens mirror and 39
I D 4 L
The quantity L D is called the f number of the telescope written f so 2 F 4 I f Notice that the image plane ux does not depend on the size of the telescope but only on its f . Photographers are familiar with this fact a light meter dictates an f and shutter speed independent of the focal length of the lens. The Palomar 20000 is a slow telescope about f 3 5 that is f 3 5. Modern telescopes are often between f 1 75 and state of the art f 1 25. That is they are short focal length and wide aperture. If image ux depends only on f number why not make tiny little fast f 1 25 telescopes Palomar in your pocket or Keck in your khakis Answers For bright objects e.g. planets we want the large image scale to get more pixels resolution elements on the detector across the image. For faint objects e.g. distant galaxies we are photon starved and care not so much about energy per unit area on the focal plane as about total number of photons from the object per time which scales as D2.
40
focal plane
eyepiece D
p eyeball
angle
D is the aperture L is the focal length of the objective lens d is the exit pupil l is the eyepiece focal length
The basic idea is that the objective makes an image at the focal plane and the eyepiece is simply a magnifying glass for looking at that image. From the gure you can derive
magni cation
L l
Notice that it is useless to have d bigger than p the pupil size of the human eye no greater than about 9 mm when dark adapted because the extra rays simply do not enter the eye Tracing the rays back the maximum useful size of D is D Md Mp Bigger than this does not make a brighter image Liouvilles theorem again. Smaller than this does make a dimmer image. How can this be true without violating Liouville Answer for ray bundles of diameter smaller than p the eye is no longer a good approximation to a surface brightness meter it is arti cially stopped down. In bright light of course the eye naturally stops down on its own when the pupil contracts. 41
I B T
The reason for this involves both some statistical mechanics and some quan tum mechanics. Roughly it is that the electromagnetic degrees of freedom of the system come to the same thermal equilibrium as the mechanical motion of particles degrees of freedom each one getting a mean energy of
1 kT 2
where T is the temperature and k 1 3806 10 16 erg K is Boltzmans constant really just a unit conversion to go from the historical but arbitrary temperature unit of degrees to the natural temperature unit of ergs. The huge importance of the Planck spectrum in astrophysics is because many things we see are perfectly or very nearly in thermal equilibrium. So the Planck spectrum is almost the universal spectrum of astrophysics.
Example 2 Looking at the dark night sky we are actually looking at the
42
mA gh kT
The Boltzman factor e E kT is actually the classical approximation valid for the atmosphere problem to a deeper quantum mechanical result which we now describe. Remember phase space density the number of particles per unit d3x d3p In quantum mechanics there is a natural unit a kind of quantum for phase space volume and it is Plancks constant cubed
h3
The occupation number of a quantum system is the number of particles in one quantum unit of phase space or in terms of the phase space density N occupation number N h3 The deep result on thermal systems is that they have a universal mean oc cupation number given by 1 occupation number eE kT 1 Notice that if E kT the 1 is negligible and we recover the classical result. There are two fundamentally di erent kinds of particles in the universe Fermions obey Pauli exclusion principle use 1 above includes protons neutrons electrons Bosons like to clump can correspond to classical wave elds use 1 above include photons gravitons Now lets put the pieces together using the result in 3.3.1 that relates N and I
B T I
3 1 h 2 c2 eE kT 1
Where did the extra factor of 2 come from in the last two equations It is the two polarizations of photons. Each one separately gets the occupation
number eE kT 1 1 and I counts both of them As before Section 3.3.2 we have the option of accounting for the intensity on a per wavelength instead of per frequency basis
B
2 B c
44
cB 2
Since wavelength is easier to measure for light than frequency one more frequently sees B and I in observational contexts. A graph of BT shows that hotter T s indeed give both overall greater intensity everywhere and peak at shorter wavelengths.
2000 K
B (T)
3 4 in m
45
Notice that this result does not contain Plancks constant. It was originally derived by assuming a mean energy E kT the classical equipartition value for the energy of an electromagnetic wave 1 2 kT in each of two polarizations. A plot of the Planck function over many decades of frequency I and tem perature shows clearly the asymptotic regimes see gure on next page. You can see why things are called red hot when the Wien limit is just barely poking into the red from the infrared then white hot when the peak of B T is at the eyes maximum visibility about 6000 K then blue hot when the eyes sensitivity is completely in the Rayleigh Jeans regime. There is nothing hotter than blue hot because higher temperature just increases the Rayleigh Jeans spectrums amplitude linearly with T without changing the functional form of its 2 dependence on frequency.
46
(cm)
106 104 10 2 1 10
2
10 6
10 10 10 12 B 1045
108
1010
1012
v(Hz) 1014
1016
10
18
10 20
10 22
10 8 10 6 104 10 3 10 2 10 1 10 5 107
1 104 108 10
12
10 16 1020 104 B 106 104 10 2 1 10 106 10 8 1010 1012 1014 1016 1018 v(Hz)
10 20 1022
10 10
10 4 10 6 108 (cm)
10 10 10 12
The peak of the black body spectrum is found letting x follows d x3 3x2ex 1 x3ex 0 dx ex 1 ex 12 x 31 e x
h kT as
hmax 2 82kT
47
or
shifts its peak a bit relative to B so at the maximum of B you can check this h c kT 4 97
max
or
maxT
The spectral sensitivity of the eye is di erent to bright light cones than to dim light rods. Bright sensitivity peaks at 5600 Ayellow with an e ective width of about 1100 A. Note close similarity to the V band in UBV Dim light sensitivity peaks at about 5000 Agreen with about the same e ective width see the gure on the next page.
T T
Not coincidence that we evolved to put our peak sensitivity at peak of solar output Dim caves moonlight etc.
48
1.0 0.9 Relative Visibility 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 350 400 violet 450 blue 500 green 550 (nm) 600 650 red 700 750 red looks "black" in dim light rods "Scotopic" dim light cones "Photopic" bright light
yellow orange
1 I d d 4 I d c c 4 2h 3 c2 d c eh kT 1 4 2h kT 4 1 x3 dx c c2 h 0 ex 1
ZZ Z Z Z
where x h kT . The integral is just a number . You can do it numerically or look it up in a table of integrals. Some day you should learn how to do it by contour integration. The answer is 4 15 so
5 k4 T 4 aT 4 u 815 h3c3
49
where a is called the radiation density constant and has the value
a 7 56464 10
15 erg cm 3 K 4
Next What is the emergent ux from a black body surface That is how many ergs per second does it emit per unit area In 3.3.5 we already obtained the result F B relating ux and brightness so
F
The physical constant has the value
c u from above I d 4 ac T 4 T 4 4
Z
4r2F
4r2 T 4
Example 2 How many watts of total luminous power are emitted by the
standard candle that de nes the lumen 1 60 cm2 at 2044 K
L A T4
In a real candle many times this much power perhaps 100 is carried o convectively by hot air. Note that at T 2044 K melting point of platinum
max
0 290 cm K 2044 K
14000 A infrared
So the candle is extremely red we are seeing only its Wien limit tail. Lumens and lux are so called photometric apparent units which therefore implicitly include the human eyes sensitivity as opposed to ra diometric units which are in absolute c.g.s. At more visible colors there are therefore many more lumens per watt than the value 2 that we got above for the dull red glow of melting platinum. At 5550 A there is a peak of 680 lumens per watt. The lumens per watt conversion at other wavelengths is this value times a standardized form of the sensitivity curve shown above 3.5.3. Light bulbs are often labeled with their lumens value as well as their wattage. For example a typical 100W bulb may yield 1710 lumens for 750 hours.
3.5.5 Color temperature and brightness temperature Color Temperature. Often a spectrum is measured to have a shape more or
less of blackbody form but not necessarily of the proper absolute value. For example by measuring F from an unresolved source we cannot nd I unless we know the distance to the source and its physical size. By tting the data to a blackbody curve without regard to vertical scale a color temperature Tc is obtained. Often the tting procedure is nothing more than estimating the peak of the spectrum and applying Wiens displacement law to nd a temperature. Rybicki and Lightman For example the color temperature of moonlight is very nearly the same 51
as that of sunlight. Long exposure color pictures by moonlight appear in nor mal colors despite the inability of your eyes cones to function at moonlight levels. Brightness Temperature. The Planck spectrum is monotonic with temperature at each and every frequency see previous gure. Therefore a measurement of I at any single frequency determines a temperature called the brightness temperature. Of course identifying this temperature with a real physical temperature of the emitting object depends on knowing the the rest of the spectrum also ts a black body shape. The Planck spectrum by de nition has the same brightness temperature at all frequencies. The brightness temperature of the night sky is 200 300 K where the atmosphere is opaque and as low as 3 K cosmic microwave background temperature in atmospheric windows of transparency.
Optical window transmission 1 0 0.1 nm 1 nm Atmosphere is opaque 2.2 3 - 5 8 - 12
10 nm 100 nm 1 m 10 m 100 m
wavelength
Note the infrared windows at 2 2 3 5 8 12 as well as the large optical and radio windows. UV and X ray astronomy can only be done from space.
the actual surface area is 4R2 so the average insolation as it is called is 1 4 the solar constant 1 1400 w m2 3 5 105 erg s 1 cm 2 F ins 4 The outgoing infrared ux at temperature T is T 4 so
5 67 10 5 erg cm 2 s 1 K
T 4 3 5 105 erg s 1 cm
280 K 7 C 44 F chilly
The reason the Earth is actually warmer than this is because of the greenhouse e ect. Atmospheric gasses absorb the IR emitted from the ground and re radiate it to space. You might think that this is a wash but not so as the following gure shows
I IR I greenhouse gases H2O, CO 2
visible
You see that the ground must radiate twice the incident ux I to stay in equilibrium. Thus
4 5 67 10 5 TK
IR
TK
The reason it is not this hot is that the greenhouse is not totally absorb ing note the partial windows between 8 and 20 in the gure in Section 3.5.5. You can see why there are environmental worries about increasing the greenhouse gasses however 53
IR
2I
Keep in mind that the above calculations are idealized because clouds re ect sunlight directly and because transport of heat from equatorial to polar regions by weather and ocean currents is a big e ect. It must still be true however that net solar ux in equals total IR ux out when averaged over time and latitude.
Arcturus Boo Aldebraran Tau Strong titanium oxide Antares Sco and some neutral Betelgeuse Ori calcium
54
Here we are seeing the Rayleigh Jeans continua ux proportional to temperature with relatively small and few absorption features. Compare shape of curves in the gure in Section 3.5.2
we clearly see the peak here of the Planck function around 4500 A
Here we are on the Wien tail with lots of messy absorption lines and molecular bands.
55
Within each spectral class there are ner steps B0 B1 B2 B9 A0 A1 A9. etc. B5 is a typical B star. The Sun is a typical G star that is G5. When astronomers nally knew the distances to enough stars to get their absolute magnitudes or luminosities and plotted these against their spectral types or color temperatures they got a big surprise
(bright) 10 Rigel 5 Absolute magnitude or Luminosity Deneb Betelgeuse Antares
+5
Barnard's star
The stars are not randomly scattered but ordered in some weird way Uh oh call in the astrophysicists This kind of diagram is called a Hertzsprung Russell or H R diagram. Ejnar Hertzsprung discovered it in 1905 and Henry Morris Russell re discovered it nearly 10 years later. News travelled slow then. There were very few astronomers in the world and not many interna tional conferences. 56
B e-
57
Radiation is less intense almost unpolarized very at spectrum in radio thermal in X ray Planck like.
typical X-ray: log I typical radio:
0.1
e/0
2.5
log
Usually associated with hot gas X rays from clusters of galaxies HII regions surface of sun.
Bremsstrahlung is usually associated with hot gas X rays from clusters of galaxies HII regions or the surface of the Sun e.g. Synchrotron radiations and thermal bremsstrahlung are two of the com mon continuum emission mechanisms of gas. Solids e.g. surface of the Earth are di erent.
Spectrum
Seen from hot gas may see lines from many types of atom simultaneously. Molecules also make lines.
There are also absorption spectra radiation absorbed by cooler material lines or continuum the inverse process.
59
4 Classical Dynamics
4.1 Newtonian gravity
4.1.1 Basic law of attraction
Two point masses with mass M1 and M2 lying at r1 and r2 attract one another. M1 feels a force from M2
F 12
where r12
force on M1 from M2
GM1 M2 r 12 3 r12
r1 r2
Coordinate Origin
r2
F 21
where r21
force on M2 from M1
GM2M1 r 21 3 r21
GM1 M2 r 12 3 r12
F 12 r12
and jr21j jr12j. Thus the gravitational forces are equal and opposite. This is in accordance with Newtons 3rd law and ensures that the composite system of M1 M2 does not suddenly start moving as a whole and violating Newtons 1st law. Notice also that the gravitational force at M1 from M2 is directed exactly at M2 . It is a central force F 12 is parallel to r 12 . This remark causes us to digress and mention 60
OK A A
NOT OK!
(a)
(b)
If b were possible then we could take the two particles A and B nail one to the hub of a wheel the other to the center and watch as the wheel accelerated up to in nite angular velocity. 61
Thus while one often hears about the central force nature of gravita tion this is actually a property of any action at a distance classical force law. You might wonder about the magnetic force between two moving par ticles in electromagnetism. This does not seem to be a central force. The explanation is that the electromagnetic eld itself carries momentum and angular momentum. Classical electromagnetism cannot be written as a pure action at a distance theory. That is there is more to Maxwells equations than just the Coulomb law.
F 12
force at M1 due to M2
r2 r3
is in
where the sum excludes any mass exactly at r do not include the in nite self energy. 62
M2r2
jr1 r2j3
1 2
These apply for example to binary stars at large separations but small enough that the tidal e ects of other stars in the galaxy can be neglected. Or the orbit of the Earth around the Sun to the extent that the e ects of the Moon and other planets can be ignored.
GM1M2 r jr r j3 2
r1
M1r1 M2r2 0
Let the center of mass of the system be at R then by the de nition of center of mass as the mass weighted average position 1 r 1 M2 r 2 R M M1 M2 63
we get
M1 M2R 0 or
Thus the location of the center of mass of the system is unaccelerated since there is no external force on the system. It follows that the center of mass moves at a constant velocity R v constant R ddt R R0 v 0 t This relation expresses the Conservation of Linear Momentum. R0 and V 0 are the rst 6 constants of motion. The next logical thing to do is to change to center of mass coordinates r01 r02 that is to coordinates relative to the center of mass.
M1 r1 r1 R r2 O CoM r2 M2
r1 r2 R
Substituting these into the equations of motion we get GM1 M2 r0 r0 M1r01 jr0 r0 j3 1 2
1 2
64
M2r02
r01
Note that the equations preserve exactly the form they had before We could have just written down the answer without calculation. We can arbitrarily choose coordinates relative to the CoM because the CoM frame is a good inertial frame. Any change of origin R0 and or change of velocity V 0 is allowed by so called Gallilean invariance. So let us choose to do this coordinates are henceforth measured relative to the CoM and we drop the primes. In CoM system GM1 M2 r r Mr
1 1
M2r2
jr1 r2j3
1 2
with M1r1 M2r2 0 by de nition. We can now construct further integrals of the motion. Consider the total angular momentum about the CoM L M1r1 r1 M2r2 r2. Then dL M1 r1 r1 r1 r1 M2 r2 r2 r2 r2 dt M1r1 r1 M2r2 r2 Substitute the equations of motion for M1r1 and M2r2 then dL GM1M2 fr r r r r r g dt jr1 r2j3 1 1 2 2 2 1 GM1M2 r r r r jr r j3 1 2 2 1 0
1 2
GM1 M2 r jr r j3 2
r 1
constant. 65
That is the angular momentum vector is xed in direction and magnitude because no external torques are acting.
L M1 r1 r1 CoM r2 r2 M2
The angular momentum provides 3 more integrals of the motion so we are now up to 9. We can also look at the total energy of the system
dE dt
Now
r 2j
d jr dt 1
Therefore
r2j
r1
dE dt
r1 M1r1
0
r 2
GM1 M2 r r2 M2r2 jr 2 1 r 2 j3
using the equations of motion. Therefore the total energy of the system
E constant
66
because no work is being done on systems by an external force. The energy is another integral of the motion each one equivalent in counting to one initial condition. These 10 integrals R0 v0 L E are general to N bodies not just 2 as we will see later. To completely specify the system we need 12 constants of integration 10 speci ed 2 remain. These turn out to be particular to the special case of 2 bodies. That is they are more like initial conditions than conservation laws.
M1 r 1 M2r2
and by de nition We showed that
jr1 r2j3
r2 r1
M1r1 M2r2 0
constant vector
E constant scalar
We now want to completely solve the system for r1t and r2t. 67
It is convenient to reduce the system to simpler form by working in terms of the separation vector of the two bodies
r r1 r2
M1
r1
CoM
r2
M2
Then consider equation 1 divided by M1 minus equation 2 divided by M2 2 r r1 r2 GjrM1 rM 3 j r1 r2
1 2
GM r r3 where M M1 M2 is the total mass of the system. This equation says that body 1 moves as if it was of very small mass being attracted by mass M1 M2 at body 2 the separation vector is accelerated by the total mass of the system. The fact that the 2 body problem reduces to the equation of the one body problem test mass in the central force eld of a ctitious mass M is an amazing and nonobvious result With the relations
or
r r1 r2 M1r1 M2r2
2 r1 M Mr
r2
68
M1 r M
Notice the appearance of a quantity M1M2 M1 M2 on the right hand sides where we expect one over a mass. The quantity is called the reduced mass. Notice that is smaller than both M1 and M2 . The angular momentum equation M r r M1M2 L has an important immediate consequence. Since r r r r r r 0 that is the vector cross product of two vectors is perpendicular to both of them r and r are both perpendicular to L i.e. all the motion occurs in a plane perpendicular to L remember that r1 and r 2 are proportional to r hence parallel to r. Adopting coordinates in the plane of the orbit we can write in Cartesian coordinates
Note that the signs are opposite the masses must be on opposite sides of the CoM. Eliminating r1 and r2 in favor of r we nd that rt obeys GM r r r3 r r MM L 1 M2 1 r r GM M E 2 r MM
1 2
L r r
The momentum equation becomes
0 0 L x y 0 x y 0
xy yx
M
L L M1M2
69
M M1M2 E
E
In principle the above two equations are the coupled di erential equations for the two unknowns xt and yt. It is simpler however to exploit the planar geometry in polar coordinates r and .
radial direction ^ r
r r
r r r r rr
r2 r2 2
70
. r
r er
r e
dA t
In interval t to t dt the vector between the objects sweeps at area dA 1 r rd dA 2 Therefore the rate of change of area is dA 1 r2 dt 2 which by the rst equation becomes dA 1 M L L constant. dt 2 M1M2 2 71
Thus it is a simple consequence of the Law of momentum conservation that the rate of sweeping out of area is a constant. This is Keplers Second Law 1609. Kepler discovered it empirically he had no idea what physics was behind it. Kepler was Tychos graduate student in modern terms but Tycho would never let him look at all the data. Tycho died in 1601 and Kepler got the data then. Even today many observers would rather take their data to the grave than let theorists analyze them. Or so it seems to theorists Proceeding we can get one equation for rt by eliminating from these equations. Then the equations can be rewritten 2 dr M E 2GM M 2 L2 2 M1 M2 r M1M2 r2 dt M L d dt M1M2 r2
M1M2 M
This is a rst order di erential equation that can be integrated Z Zr L dr d 0 h L 2 i1 2 r2 2E 2GM 2 r r 72
r0
L 2 GM
L2 GM2
2EL 1 GM 23
2
0
1 1 cos 0 r0
the plus or minus is thrown in gratuitously because it corresponds to 0 an unknown constant changing by so we can write the result with the more convenient sign 1 1 1 cos 0 r r This is the equation of a conic section. There are three special cases which we now consider.
0
Case of
1 0 for all so the separation between the two masses If 1 then r remains nite the objects remain bound with
r0 r r0 1 1
73
x y
r cos 0 r sin 0
and eliminate and r the equation of the orbit becomes 1 1 1 x x2 y21 2 r x2 y21 2
0
a 2 y2 a b2
a
a
r0 x
x = a(1 ), r = r0 /(1 + ),
y=0 =
x = a(1 + ), y = 0 r = r0 /(1 ), = 0 +
The axial ratio of the ellipse is b a 1 21 2. The quantitiy r0 is called the semi latus rectum. Note that 1 implies from the de nition
21
We can also see that the orbit is periodic and closed. Let us nd the period 1 1 cos 1 0 r r0 d L dt r2 are the equations governing t we can write this as a single equation for t Zt 2 Z r0 d dt L 1 cos 2 Now in one repeating orbit
0
increases from 2 t increases from t t period so 2 Z 2 r0 d L 0 1 cos 2 2dt 1 t2 The substitution t tan 1 2 so d 1 t2 cos 1 t2 renders this integral tractable. Then 2 r0 2 L 1 23 2 Eliminating r0 and in terms of physical variables L GMr01 2 and r0 a1 2 a the semi major axis 32 a 2 GM 1 2 Note incidentally you can check that a depends only on the total energy a GM1M2 2E while depends on both the energy and the angular momentum. For bound orbits we have now proved Keplers laws
1. The planets move in ellipses with the sun the center of mass at one focus 1 1 1 cos 0 r r
0
75
with
1
0 is circle.
2. A line from the Sun to a planet sweeps out equal areas in equal times
dA dt
1 L 2
Note this also works for unbound orbits. 3. The square of the period of revolution is proportional to the cube of the semi major axis 2 a3 2 GM or 2 GM a3 Note this involves the total mass M
M1 M2.
Case of
1
2EL 1 GM 23
2
we see that E 0 so the total energy of system is positive. This will imply that the orbit is unbound 1 r 1 1 r0 cos 0
is a hyperbola x r cos 0 y r sin 0 gives again 1 1 1 x x2 y21 2 r x2 y21 2 1 a 2 1 b 2 r011 2 b a 2 11 2
x
a2 r0
a2
y2 b2
76
r0 M1
M2 x
Case of
1
0 i.e. total energy of system is
From the de nition of 1E zero. In this case the orbit is a parabola
1 1 1 cos 0 r r0 2 2r x y 2 r0 0
y
r0 M1 r /2 0
M2 x
77
78
The 6 orbital elements are de ned as 2 shape parameters semimajor axis eccentricity inclination longitude of the ascending node
a e i
half the length of AA0 1 b2 a21 2 if b is the semiminor axis angle 6 QNB between the orbital plane and the ecliptic plane angle 6 SN from rst point of Aries to the line of nodes N 0N . Ascending means pick the node where the planet crosses the ecliptic from south below to north above.
angle 6 NSA1 from the ascending node measured in the plane of the orbit to the perihelion point A.
1 parameter speci es orbital phase time of perihelion passage T one of the precise times that the object passes through to the point A
Note that P the period is not an orbital element since you can compute it from a by P in years a in AU 3 2 Instead of people sometimes quote the longitude of perihelion
6
SN
6
NSA1
The symbol is actually a weird form of pi but most astronomers call it pomega Note that the two summed angles are measured along two di er 79
ent planes This is not really the longitude of anything but it approximates it when the inclination i is small. Also instead of T people sometimes give the longitude L of the planet at a speci ed epoch time. Here the longitude is again a kind of phoney
L 6 N
6
NP1
In case you want to compute locations of the planets here are the values of their orbital elements called ephimerides the plural of ephemeris
Mean Longitude Mean Longitude Mean Longitude of Node of Perihelion of Epoch L
48 19051 0021495 76 40047 0071268 0.0 0 49 33 29 0013554 100 27051 0098631 113 39055 0088533 74 00021 0041002 131 47002 0060528 110 17049 007 77 270 22 0002855 131 330 49 0034607 102 560 14 0045310 336 030 36 0084233 14 190 52 0071326 93 030 24 0043421 173 000 18 0057320 48 070 25 0028581 224 080 05 005 80 252 150 03 0025985 181 580 47 0028304 100 270 59 0021464 355 250 59 0078866 34 210 05 0034211 50 040 38 0089695 314 030 18 0001840 304 200 55 0019574 238 440 38 002
4.2.5 The mass of the Sun and the masses of binary stars
We have seen that the sidereal period of a planet major axis of its orbit a by 2 4 a3 2 GM where is related to the semi
M M
mplanet
This formula has been applied to calculate GM on the last column of the following table. Observed Inferred Sidereal Semi major period axis of orbit GM 26 Planet days AU 10 cm3 s 1 Mercury 87.969 0.387099 1.32714 Venus 224.701 0.723332 1.32713 Earth 365.256 1.000000 1.32713 Mars 686.980 1.523691 1.32712 Jupiter 4332.589 5.202803 1.32839 Saturn 10759.22 9.53884 1.32750 Uranus 30685.4 19.1819 1.32715 Neptune 60189 30.0578 1.32723 Pluto 90465 39.44 1.32727 81
GM
This is known very accurately to 1 part in 106. Amazingly G itself from laboratory experiments is only known to about 1 part in 104 as 6 670 0 004 10 8 cm3 s 1 g 1. So our knowledge of the mass of the Sun is limited by laboratory physics not astronomical observations
GM GM
MJ ME
1 1
where ME is the mass of the Earth. Jupiter has about an 0.1 e ect on GM . Thus MJ 0 1 of solar mass. Furthermore
MJ 0 000949M 1 89 1030 g
This is fairly inaccurate 1 part in 103 and it is a good illustration of an important principle normally we measure planet masses by their e ects on the satellites or one another. Arti cial satellites e.g. Voyager II at Uranus are also good This works for our Sun but what about other stars Their planets are too faint there are no low mass test particles that we can see around these 82
stars. But 50 of all stars are in multiple star systems. So we can use the relative motions of binary stars to get stellar masses. The periods of binary stars vary typically from a few hours e.g. 6h for Mirzam alias Ursa Majoris to hundreds of years e.g. 700 yr for 61 Cygni. Some are even shorter but for these it is likely that the stars are highly distorted by tidal e ects and their masses will be subject to systematic errors e.g. semidetached or contact binaries. A variety of data may be available for a binary star. We may see the motion of one star around the other on the sky which is rare or we may see only variations in the velocities of the stars as seen from Doppler shifts in their spectra this is more common. Binary stars may be single lined see light from only one star or double lined. Consider for example the second case a double lined spectroscopic binary star. Here we see the two stars spectra superimposed and we can measure the radial velocities for both stars but the stars are too close together for us to see their orbits by measuring the changing positions of the stars relative to one another. Then what we see are star velocities like
velocity, v
.5
time, t /
1.5
83
Immediately we see that we have three observables the orbital period the peak velocity of the less massive star 1 K1 the more massive star moves more slowly the peak velocity of star 2 K2 The shape of the curves is here sinusoidal which implies circular orbits 0. Other shapes are possible for O 1. What can we get from these data To simplify things suppose the stars are in a circular orbit 0 not necessary we could t from the velocity data if we wanted to but this makes things harder. Then our orbit equations are
2
But the period is related to the separation of the stars r the same as the semimajor axis for a circular orbit by
2
42r3 GM
and the speeds of the stars in their orbits are v1 r1 and v2 r2. If the orbital plane is parallel to the line of sight the peak velocities we would see for the stars are r1 and r2. In general however this will not be the case see gure on the next page. At inclination angle i i angle between line of sight and normal to orbital plane 84
K1 K2
r r1 r2
Using the previous formula for gives a total mass
2 and eliminating r
M1 M2
The center of the mass provides the constraint M1r1 ratio is M2 r1 K1 M1 r2 K2 Putting these together
M1 sin3 i M2 sin3 i
Here the right hand sides are completely known so we infer the left hand sides. These data only give masses if we know the inclination of the orbit i. But in general we will not know this so our data only measure a combination of masses and i. For one special group of stars we do get M1 and M2 directly from eclipsing binaries where the light curve shows an eclipse see gure.
1.2 secondary eclipse primary eclipse
1 .8 .6 .4 .2 0 0
.5
1 time, t /
1.5
Here the stars are eclipsing one another . This scenario is only possible if i 90 provided that the stellar radii separation which we can check in the above equations if this is not true the stars are so distorted that we cannot use this method anyway. Thus sin i 1. Errors in i have little e ect on the masses derived because sin i changes slowly at i 90 . In fact the best data on masses of stars comes from eclipsing spectroscopic double lined binaries even though these are hard to nd. An alternative approach is to measure many noneclipsing binaries assume i to be selected at random and use statistics to derive the distribution of masses M1 and M2. 86
r1
CoM
r2
M2 v2
0 0
Let M1 be the mass that explodes. Typically it is the more massive star that explodes so M1 M2. When it explodes it quickly and spherically say blows o most of its mass so that its new mass M10 is
M10 M1 M
and we have
87
v1 M1
M2
old CoM
new CoM
v2
A rst important e ect is that the remaining binary is not at rest. Lets calculate its new CoM velocity vc
MM2 M1M1 M2 M v2 Notice that as M M1 all star explosion vc v2 as it must. Typical values might be M1 10M M2 5M M 8 5M leaving a 1 5M neutron star for M10 so
8 5 5 v 0 65v 2 10 6 5 2 88
For close binaries v2 can be hundreds of kilometers per second so the system really takes o In fact we need to check whether the binary remains bound at all The easiest way to do this is to see if its total energy in the new CoM is positive or negative just after the explosion 0 total 1 0 kinetic 1 total energy potential B C B C 0 E ( kinetic A ( energy A in new CoM energy energy of CoM 1 1 0 0 v2 1 M v2 0 M v 2 GM1 M2 M M 2 c 2 11 2 22 2 1 r Recall that Section 4.2.2 the separation vector r and the relative velocity v2 v1 satisfy Keplers Laws with the total mass. So for circular orbits
GM1 M2 v 2 r
v12
Substituting this and v c previously obtained and also eliminating M10 in favor of M1 M and v1 in favor of v2M2 M1 we get M2v2 2 1 1 0 2 1 M E M M 1 M 2 v2 1 M2 M 2 M 2 2 1 2 M2 2 MM M 2 v2 1 M M2 1 M v2 M M M M M1 M2 1 1 2 1 Now a fairly amazing algebraic simpli cation I used the Simplify function in Mathematica on the computer gives 1 2 M1 M M1 M2 M E0 M v 2 2 2 M12M1 M2 M 1 M2 2M
Notice that all the terms are positive since M M1 except the last one. We see that a condition for E to be positive unbinding the binary is M 1 M M 2 2 1 89
that is loss of more than exactly half the total mass of the system. In the previous numerical example we have 85 1 10 5 2
so the neutron star is unbound and ies away at high velocity. In fact pulsars which are neutron stars are often seen to be leaving the galactic plane where they are formed at high velocity for just this reason. You might wonder if there isnt a simpler way to derive the above half of total mass result. There is if you are comfortable with expressing the energy of a 2 body system in terms of the reduced mass as in above section 4.2.2. Then as Mike Lecar pointed out to me you can write the total energy of a moving or stationary binary system as 1 GM 1 2 2 E 2 MvCoM 2 v r where M is total mass vCoM is the center of mass velocity and v is the relative velocity that is jrj. The rst term has no e ect on whether the binary is bound or unbound. Rather it is just a question of whether the second term is positive or negative. For an initial circular orbit we have 1 1 GM 2 GM v2 GM so that v r 2 r 2 r After the supernova v2 is instantaneously the same while and M change. You can see that losing more than half of M causes v2 2 GM r to change sign become unbound.
of the other body. For the Earth Moon system e.g. this causes the surface of the oceans and to some extent the solid Earth as well to be deformed into the familiar tidal shape
least gravitational attraction greatest gravitational attraction
Earth
Moon
Our goal is to understand the above picture quantitatively. The surface of the ocean is an equipotential because if it were not that is if there were a potential gradient along the surface there would be a force causing the water to ow downhill until it lled up the potential valley. But is it a purely gravitational potential we must consider No because the stationary frame in which things can come to equilibrium is the rotat ing frame in which the Earth and Moon are xed. We are here assuming a circular orbit. Things would be more complicated if the orbit were sig ni cantly eccentric. So there is also a centrifugal force and corresponding centrifugal potential. If the orbital angular velocity is with M2 2 GM1 3 R then the centrifugal acceleration at a position r with respect to the CoM is r so the centrifugal potential whose negative gradient is the 1 j r j2 or in the orbital plane 1 2 r2 . The total potential acceleration is 2 2 in the rotating frame is thus GM1 GM2 1 2r2 jr r1j jr r2j 2 91
The gure shows an example of such a potential. Notice that the downhill gradient tries to make a test particle either fall into one of the potential wells of the masses or be ung o to in nity by centrifugal force. We will come back to this general case below in Section 4.3.4.
D2 R2 r2 2Rr cos
and then expand in the small ratio r R. 92
sub-lunar point
jr r2j
GM2
We need to do the binomial expansion to second order to consistently pick up all terms of order a2 R2 . This gives GM2 1 a cos 1 3 cos2 1 a2 O a
3 R R 2 R2 R If you know about multipoles you will see the Legendre polynomials in cos lurking here. Note that the term a R cos actually varies linearly in the z direction from M1 to M2 since z a cos so the gradient of this part of the potential is a constant force.
z
Now for the centrifugal potential term we again apply the law of cosines 1 2 r2 1 2 M2 R 2 a2 2 M2 R a cos 2 2 M1 M2 M1 M2
M1 a Earth
Moon M2
M2 R M1+ M2
CoM
So collecting the three terms in the overall potential and substituting for 2 formula in 4.3
2 a2 GM a 1 1 GM2 1 R cos 2 3 cos 1 R2 a R 1 GM1 M2 M2
2 R2 a2 2 M2
Ra cos 2 R3 M1 M2 M1 M2 Look carefully and you will see that the term in a R cos exactly cancels out. This is not coincidence it is because the constant force term is exactly canceled by the centrifugal force that keeps the bodies in a circular orbit. Also note that there are terms with no a or dependence Since these are just constants they produce no gradients and can be ignored. So we get combining the remaining terms GM1 1 Ga2
3M cos2 M constant a 2 1 a 2 R3 This is the local tidal potential near the Earth. To get a better feeling for its shape let us expand a in terms of height h above the radius of the Earth mean sea level 2 h h a R h a2 R2 a 1 R1 1 R 1 R
94
Here we have used the binomial expansion to get a2 and a 1 to rst order in h R . Now to this order 2 GM h 1 GR 2 h 1 2 h R 1 R 2 R3 1 R 3M2 cos M1 constant
We can now absorb into the constant all terms that depend on neither h nor . Doing this and with some rearranging of terms putting all multipliers of cos together and all remaining multipliers of h together we get 4 GM R M 3 h 1 2 R4 2 h cos M R3 2 R constant 2 h 1 R4 R 1 M2 R 3 3 2 g h 2 M R R cos constant 1 Here the approximate equality means we are neglecting terms that are smaller than the dominant terms by either factors of R R or h R both being assumed small. Note that g is the acceleration of gravity at the surface of the Earth. The surface of the ocean in the gure at the beginning of 4.2 is an equipotential so it must have h constant implying 3 M2 R 3 R cos2 constant h 2 M1 R Since cos2 varies between 1 front and back of Earth and zero sides of Earth the other factors give the height of the tide high minus low. Putting in M2 M1 1 81 R 6400 km R 380000 km we get h 54 cm. The Sun also generates a signi cant tide that to lowest order just adds with that of the moon but with a di erent origin for the direction pointing to the Sun. For the Sun 3 3 6400 km h 2 332000 1 5 108 km 6400 km 25 cm 95
Notice that we never assumed that M1 M2 was either 1 or 1 so our formulas apply both for the Sun and for the Moon. It is a numerical coin cidence that the tides they raise are of the same order though some have speculated that the resulting more variable tides produced could have been somehow necessary to mix the oceans in a way that furthered the evolution of life or its emergence from the sea.
36 0 /d ay
13 /day
2. The drag slows down the rotation of the Earth. Since angular momentum must be conserved as the Earth loses angular mo mentum it must be gained by the Moons orbital motion How does this 96
come about The non symmetric bulge angle in above picture causes a gravitational torque back on the Moon which transfers exactly the right amount of angular momentum. This must work out exactly because angu lar momentum is truly conserved Note that energy is not here conserved because friction is present between the ocean and the rotating Earth. Increasing the Moons orbital angular momentum causes its orbit to grad ually move outward and the length of the month gets longer. In fact we know from growth scales in fossil corals in which annual and monthly patterns can be discerned that the day and month were both once substantially shorter than now Earth has slowed down and Moon has moved farther away as we predict. Where will this process end As the Earths rotation slows down the Earth will ultimately come to rotate once per month the then length of the month and keep one face always towards the Moon. In this state called tidally locked there is no drag on the tidal bulge it will point exactly at the Moon as Earth and Moon rotate. In fact this corotation the current state the Moon is in with respect to the Earth. What will the ultimate length of the day and month be Let 1 2 1 day 2 2 1 month f 2 ultimate day and month. Then conservation of angular momentum gives 1 Earth 2 Moon angular momentum now 1 M2 I11 I22 MM 2R2 angular momentum then 1 M2 I1f I2f MM f R2 f
while Keplers law gives the future radius R0 of the Moons orbit 2 3 Rf 2 R f 97
to good approximation since M1 81 3M2 we have M1M2 M M2. Also we can neglect the spin angular momentum of the Moon now the Earths alone being so much larger and we can neglect both spins in the nal state. If you doubt these approximations you can substitute back the answer we will get and check retroactively that they are justi ed. Also we will use the moment of inertia of a uniform sphere even though this is not quite a true assumption 2 M R2 I1 5 1 1 Then we have 2 M R2 M R2 M R2 2 2 2 f f 5 1 1 1 yielding
M22
R2
1 3 2 f
2 f
2
So the nal length of day and month will be 28 1 99 54 days and the nal moons orbital radius will be its current value times 1 992 3 or about 590000 km. The current rate of lengthening of the day is about 20 per 109 years so it will take 1010 yr for the process to go to completion. By then the Sun will have swollen to a giant and incinerated Earth and Moon anyway. Incidently although our speci c example has been the Earth Moon sys tem exactly these e ects occur in close binary stars Tidal drag causes them to come into corotation tidally locked to each other. 98
Satellite: M2 B
r
M1
A z axis R
It turns out that if R is too small the satellite is torn apart by tidal forces. To see this lets calculate the gravitational acceleration at points A and B in the gure. By symmetry it must be along the z axis so 2 d GM 1 Gz gz rz dz jzj 2 R3 3M2 GM1 z z 3GM2 Gz M1 3M2 jzj3 R3 jzj3 R3 This is a restoring force only if the coe cient of z is negative. Otherwise the force is away from the center of the satellite and the satellite is torn apart starting at its surface. Putting jzj r and considering a satellite of mean density so 4 r3 M1 3 99
we get the condition for Roche stability 9 M2 4 R3 Equivalently no satellite of density can be stable inside the Roche radius 1 3 M2 Rcrit 9 4 If the parent body e.g. a planet has the same density as the satellite then M2 4R3 2 3 and we get
jr r1j jr
2
GM1
GM2
1 2 r2 r2j 2
GM1 M2
jr1 r2j3
100
A more carefully drawn version of the potential surface plot in 4.3 is the following contour plot
The equipotentials const. for the Newtonian plus centrifugal potential in the orbital plane of a binary star system with a circular orbit. For the case shown here the stars have a mass ratio M1 M2 10 1. The equipotentials are labelled by their values of measured in units of GM1 M2 R where R is the separation of the centers of mass of the two stars. The innermost equipotential shown is the Roche lobe of each star. Inside each Roche lobe but outside the stellar surface the potential is dominated by the Coulomb 1 r eld of the star so the equipotentials are nearly spheres. The potential has local stationary points r 0 called Lagrange points at the locations marked LJ .
The picture is drawn for the particular case M2 M1 0 1 but would be qualitatively similar for other mass ratios. Focus attention on the contour marked Roche lobe that de nes the saddle point between the two potential minima. Look back at the surface plot in 4.3 if you need to visualize the saddle point. In a binary star system each Roche lobe is the maximum size that its respective star can be. If due to stellar evolutionary e ects one star tries to swell up to larger than the Roche lobe it simply dumps mass through the lip of the pitcher across the saddle point onto the other star. This is called Roche lobe over ow 101
and can profoundly a ect the evolution of the binary star system. For example consider a case where M1 is initially greater than M2. Higher mass stars evolve faster so M1 may rst swell up and dump mass onto M2 possibly also losing mass to in nity as a stellar wind. Eventually M1 becomes a white dwarf star and starts to cool. Later as M2 evolves now faster because it has gained mass it may swell up and Roche over ow back onto the white dwarf a process that can be observed as an X ray source the X rays being produced as the matter crushes down on the white dwarf surface. Finally the excess mass dropped onto the white dwarf may be too much for it to support and it may collapse to a neutron star in the process blowing o part of its mass as a supernova explosion Notice that the saddle point marked L2 is a very slightly higher contour than L1. Thus a star that swells very slowly will dump preferentially onto its companion but more rapid swelling can over ow both L1 and L2 simul taneously. Matter that goes over the lip L2 is ung out to in nity in what would look like an expanding spiral trail remember the coordinate system is rotating. This is thought to be happening in so called violent mass transfer binaries of which examples are W Serpentis and Lyrae.
M2
M1 0
102
Note that the total mass M is constant. Now write the angular momentum for a circular orbit
1 2 M1M2 1 2 1 2 M M R1 2G1 2 M L R2 R2 GM 1 2 R3 M1 2 R G Now we conserve angular momentum dL 1 2 1 2 0 dt G M M1M2R1 2 M1M2R1 2 1 2 M1M2RR Solving for R and eliminating M1 in favor of M2 we get M2 M1
R 2R M M M2 1 2
12
1 2
So if the lighter star is losing mass M1 M2 then since M2 0 we get R 0 and the stars draw gradually apart. Often this turns o the mass ow since it puts M1 deeper into its Roche lobe. Or the mass ow proceeds only on the stellar evolutionary timescale that it takes M1 to keep swelling to ll the ever larger Roche lobe. Things are di erent if the heavier star is losing mass M1 M2 Then R is negative. The mass ow causes the stars to get closer which typically increases the mass ow and so on in a catastrophic instability. This could theoretically end when the stars reach equal mass. But in practice the mass ow is often so violent that hydrodynamic friction results in the stars merging.
L jrL1
rCoMj2
103
Referring to the gure in 4.3.4 you can see that jrL1 rCoMj can be signif icantly less than r the 2 body separation. A consequence is that gas that ows over L1 has only enough angular momentum to go into a circular orbit well inside the other Roche lobe if it does not impact the surface of the other star rst. If the other star is a white dwarf or neutron star so called compact objects it is small enough to allow the formation of an accretion disk of gas see gure. Once in the accretion disk gas slowly spirals into the compact object by the action of viscous forces including magnetic and turbulent viscosity.
"Normal star"
Disk DDsk L1
maintain a constant position in the rotating frame that is co orbit with the same period as the two main bodies at constant position relative to them The condition for such an equilibrium is that there be no acceleration at the chosen position that is the gradient of the potential including centrifugal term must vanish
r
Gradients vanish in general at extrema and at saddle points. In the contour plot you can see that there are 3 saddlepoints co linear with masses M1 and M2. It is easy to see that there must always be these three for any mass ratio as the following graph illustrates
Gravitational force due to M1 Negative of centrifugal force O L3 A L1 B Gravitational force due to M2 L2 x
You can see that gravitational force and negative centrifugal force must always cross at exactly three points where
Fgrav
These three points L1 L2 L3 are the rst three Lagrange points where a test mass can orbit. However since they are saddle points see contour 105
If in the contour picture is out of the page M2 and M1 orbiting counter clockwise then a particle pushed outward from L4 or L5 experiences a clockwise force and goes into a small stable clockwise orbit around L4 or L5 respectively. The best real life example of objects orbiting a stable Lagrange point are the Trojan asteroids which orbit L4 and L5 of the Sun Jupiter system. Between 1906 and 1908 four such asteroids were found the number has now increased to several hundred see gure. These asteroids are named for the heroes from Homers Iliad and are collectively called the Trojans. Those that precede Jupiter at L4 are named for the Greek heroes plus the 106
picture all three are unstable if the mass is slightly perturbed it falls into one of the potential wells or is ung o to in nity. More interesting are the two maxima of the potential labelled L4 and L5 in the contour plot. These are the so called stable Lagrange points. Looking at the contour plot or the surface plot at the beginning of 4.3 you can see that these maxima occur along the ridge line of a long banana shaped ridge between the potential well of the combined masses and the centrifugal force potential that decreases toward in nity. If a test mass is placed at L4 or L5 it will stay there. If it is perturbed slightly it will execute in the rotating frame a stable though not necessarily closed orbit that goes around L4 or L5. You might wonder how such an orbit can be stable if it is going around a potential maximum not minimum. The answer is that in the rotating frame there is a Coriolis force v so that a test particles equation of motion is actually dv r v dt
Trojan spy Hektor and those that follow Jupiter at L5 are named for the Trojan warriors along with the Greek spy Patroclus. Some of the Trojans make complicated orbits taking as long as 140 years to meander around the banana of the potential meanwhile of course orbiting the Sun every 11.86 years just like Jupiter. Abell 7th Ed. 18.3
We have not yet determined the location of L4 and L5 in the gure. The amazing fact is that independent of the mass ratio M1 M2 L4 and L5 are exactly at the vertices of equilateral triangles formed with M1 and M2. This is so remarkable that we should at least verify it for fun
107
P F1 1 F2
mass m
A M1
60
O CoM pR
60
R (1 p)R
B M2
We want to see that F 1 F 2 exactly balances the centrifugal term mr2. The magnitudes of F 1 and F 2 are
F1
GmM1 R2
Gm1 pM R2
F2
GmM2 R2
GmpM R2
32 r
sin 1 pR
108
3pR 2r
Using the previous formulas for the magnitudes F1 and F2 the force check becomes p p 3 pR G 1 p M 31 pR GmpM F sin F1 sin 1 2 2 2 2r R 2r R2 Hey it checks So now we need to check the component parallel to r which includes the centrifugal term
So
Gm1 pM R2
1 pr2
GMmr R3 r
2r2
Hmm. Is this true for all values of p Yes It is the Law of Cosines applied to the triangle BOP using cos 60 1 2
r2
N 1X 2 2 i 1 mivi
V
Newtons force law is
X X Gmimj
i j
jri rj j
mivi
rV r
De ne a quantity a bit like a moment of inertia but around a point not an axis N 1X I2 miri2 Di erentiate with respect to time twice
i
X
i
mi v i r i
110
Now here is the tricky part The virial theorem comes about because the potential energy V has a scaling relation that describes how its numerical value would change if all positions r were stretched by some factor . In particular you can see right away that
V r 1 r2
1 V r r N 1
r2
V
X
i
r rV r r
i
Thus
I V 2T
This is called the time dependent virial theorem. It says roughly that if V 2T 0 the mean square size of the system must eventually be increasing while if V 2T 0 the size must eventually be decreasing. Remember that V is negative for this to make sense. It also follows logically enough that if a gravitating system is in equilib rium neither increasing nor decreasing steadily in size it must have
hV 2T i 0
or
2hT i
hV i
Here the angle brackets denote the long time average. To see this formally average the time dependent theorem over a long time T ZT ZT 1 1 I T I 0 1 hV 2T i T 0 V 2T dt T 0 I dt T 111
Now if all particles remain bounded with bounded velocities for all time the de nition of an equilibrium system then I t remains bounded see its formula above and the right hand side goes to zero as T 1 thus proving the desired time averaged virial theorem.
112
cosmology. To refresh your memory about the elements here is the Periodic Table.
The numbers shown on the periodic table are the atomic numbers or num bers of protons Z in the nucleus. This of course determines the number of electrons and hence the chemistry. For nuclear processes however we also care about the total number of nucleons protons plus neutrons A in the nucleus. For all the elements from He not H through S sulfur the relationship is A 2Z within one AMU. That is there are close to equal numbers of protons and neutrons. Heavier elements get slightly richer in neutrons so A becomes somewhat larger than 2Z but not much. Iron for example has Z 26 A 56. Studying their spectra it is found that there are two distinct classes of stars called Population I and Population II usually read as Pop One and Pop Two. Pop I stars are young stars like the Sun created by 114
ongoing star formation processes within our Galaxy or in other galaxies. Pop II stars are old stars and are thought to fossilize the initial epoch of star formation after the Big Bang. When this classi cation was rst devised the connection with age was not immediately obvious so the numbering scheme I and II is perhaps illogical. A good way to remember it is to recall that a II year old child is older than a I year old child. The observational distinction between Pop I and Pop II is that Pop I stars e.g. the Sun are relatively rich in heavier elements. For astronomers heavier means anything beyond H and He in the periodic table that is Li Be B C N O etc. Chemists would consider these to be light elements Another astronomers habit of which you should be aware is that the heavy though not really heavy elements are often referred to loosely as metals even though theyre not. So to astronomers the whole periodic table is reduced to H He and metals. We know quite a lot about the elemental abundance of Pop I stars be cause we have one right in our neighborhood the Sun and also because except for H and He which escaped into space the Earth itself has in most respects the exact elemental abundances of a typical Pop I star. Here is a table and also graph note logarithmic scale of these Solar System abundances.
115
Pop I Abundances
Mass Atomic Number main isotope Relative Number Mass Fraction
H He C N O Ne Mg Si Fe
H 1 10-1 He
1 2 6 7 8 10 12 14 26
1 4 12 14 16 20 24 28 56
1 7 10 4 10 9 10 7 10 1 10 4 10 4 10 3 10
2 4 5 4 4 5 5 5
10-2 10-3
10-4 10-5 10-6 10-7 10-8 10-9 10-10 10-11 10-12 C O N Ne Na F
B
Mg Al
Si P
S Cl Ar K Sc Ca Ti V Cr Mn
Fe Ni Co Zn
Cu
Li Be 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Atomic Number
Hydrogen and Helium are primordial to the big bang in fact the Helium is produced during the rst few minutes of the big bang. All the elements from Carbon on are produced solely in stars. The combined mass fraction of metals Carbon and heavier is usually denoted Z so Pop I stars have typically Z 0 02. Dont confuse this use of the letter Z with its other use denoting atomic number. They are unrelated. Note the very low abundances of Li Be B. These are hardly produced in the big bang at all and they are in fact destroyed by stars they get cooked into heavier elements. Also note that the even elements tend to be more abundant by a factor of order 10 than their odd neighbors. This is 116
because they get built e ciently out of particles Helium nucleii that contain 2 protons 2 neutrons. So Pop I stars are made of material that has already been processed through an earlier generation of stars either Pop II stars or else older Pop I stars whose creation and destruction is a continuous process. In fact many astrophysicists study what amounts to the ecology of stars that is processes by which they form evolve are disrupted and return their now processed material to the interstellar medium for formation in a subsequent generation of stars. Since Pop II stars were as far as we know the original generation of stars you would expect them to have much lower metal abundances. Indeed that is true. Typically they might have Z 0 002 about 1 10 of the Solar abundances. Some have even smaller Z values. In our Galaxy Pop I stars are located in the disk while Pop II stars are in the so called bulge and halo. Thus we know that these parts of the Galaxy being populated with primordial stars are older.
4 MH 4 1 008 1 MHe 4 0027 M 4 032 4 0027 0 007 4MH 4 032 So 0 7 of the mass of each proton is converted to energy E mc2. In practice because of the various rules governing nuclear reactions and their probabilities the reaction is not simply
1H 1H 1H 1H 4He
That is
not
The leading superscripts are used to indicate mass number. For one thing the above does not conserve charge Instead the reaction proceeds by a series of 2 body interactions. In lower mass stars 1 5M the so called p p cycle dominates
2D e 2 D 1 H 3He 3 He 3 He 4He 1H 1 H
1H 1H
13N
H 14 N 1 H
13C 1 15O 15 N 1
decays decays
118
e
4
He
Notice that the 12C is a catalyst in this case it comes into the cycle at the top and is regenerated at the bottom. The net reaction involves only the particles shown above in boldface namely 4 1H 4He 2e 2
The s are just energy coming o . The key fact about nuclear reactions is that they are extremely temper ature sensitive. That is above a certain threshold a small increase in tem perature makes a huge increase in reaction rate. This fact makes most stars thermally very constant across a wide range of stellar masses the central temperatures of stars where the nuclear burning takes place are between 1 107 K and 2 107 K. That is this small temperature range translates into the entire range of luminosity needed to hold up against gravity both massive stars high luminosity and light stars low luminosity. A good empirical approximation for the central temperature of main sequence stars is 1 3 M 7 Tc 1 5 10 M K
exactly half the gravitational potential energy. There is no electromagnetic potential energy to speak of because of bulk charge neutrality. We can estimate these two terms as follows Z Z Gdm1dm2 GM 2 PE r12 R M kT 3N KE particleskT 2 m
p
Here M is the stars mass R its characteristic size radius say T its char acteristic central Tc say temperature. Twiddles mean equality in order of magnitude i.e. ignoring numerical constants like 2 or . Equating at twiddle accuracy P.E. and K.E. and using the empirical law for Tc given previously which was motivated by the temperature sensitivity of nuclear reactions we get a mass radius relation for stars 2 3 GMm GM m M p p R kT k15 106 K M 6 67 10 8 M cm 23 16 6 6 10 1 38 10 15 10 M 2 3 M 1 0 1011 M cm The actual solar radius is 7 1010 so our twiddle calculation is actually pretty good the neglected numerical factors cancel to near unity How do we do for other stars on the main sequence Here is some actual data
120
Mbol
12.1 10.9 9.7 8.4 6.6 4.7 2.7 0.7 11 29 46 63 76 89 10 2
MV
15.5 13.9 12.2 10.2 7.5 4.8 2.7 1.1 02 11 22 34 46 56 63
logR R 09 07 05 03 0 14 0.00 0.10 0.32 0.49 0.58 0.72 0.86 1.00 1.15 13
"actual data"
0.5 0.5
0.5
What about the luminosity of stars Can we predict that with twiddle calculations Yes but we need to know something about the opacity of stellar material that is how much resistance it gives to the outward di usion of 121
photons. I will write down the calculation even though it is beyond the scope of this course. You can study it for extra credit or intellectual curiosity. L Flux c daT 4 Di usion Energy Density Coe cient Gradient 4r2 3 dr Here is the opacity which comes from atomic physics. Let us assume that this is a constant as it very nearly is for highly ionized matter. Then neglecting constants we have
M R3 T M R
So
virial theorem
R2 T 4 R5 M R4 M 3 R M R Notice that the Rs cancel so we never had to use the mass radius relation or the empirical formula for central temperature but only the virial theorem. How good is this Since we do not at this stage know a numerical value for we cannot check the constant but only the scaling from the solar value L
6 prediction normalized here 4 2 log(L /L )
da ta
ic red
tio
0.5 2
0.5
1.5
log(M/M )
122
Pretty good So it looks like stars really do obey the laws of physics. This motivates us to do a more careful job of writing down their governing equations of stellar structure which we will do next.
and
M r
mass This is the mass interior to a radius r. It comes into the calculation of the gravitational force as a function of r. It is zero at r 0 and equal to the stars total mass M0 at the stellar surface. 123
pressure The pressure at any r is of course the total weight per square centimeter of all the overyling mass. temperature Gas and radiation are in local thermodynamic equi librium. luminosity This is the net outward total energy ow at each radius. It grows from zero to the stars total luminosity as we pass from r 0 through the stars energy generating region and thereafter as we move outward is constant at the stars total luminosity.
There are also various quantities describing the microscopic local proper ties of the gas
r
mean particle mass also called mean molecular weight This comes into the perfect gas law P kT . It is the mean counting both electrons and nucleons as particles since both contribute to pressure. opacity or mass absorption coe cient This has units of cm2 g and is the total cross sectional area for absorbing or scattering pho tons per gram of material. It controls the rate at which photons di use outward to transport the luminosity. The nuclear generation rate in ergs per cm3 per sec.
r
r
Now if you are lazy here is the good news In this course we are going to use various tricks or approximations that close the set of equations with only the rst three variables above M P . This closure is not exactly 124
correct for all cases but it will allow us to learn some interesting things about some real stars and will let us defer the whole subject of radiative transfer involving T L K and E to later courses Astronomy 145 and 150.
The pressure P r is larger than the pressure P r dr by just the weight per unit area of the material between r and r dr in the local gravitational acceleration g. If the element is of area dA we have mass of dA dr element GM r weight of r dA dr g dA dr 2 element r G rM r dr weight of element per unit area r2
dP P r dr P r G rM r dr dr r2 The minus sign is because pressure increases as r gets smaller downward direction.
So
Mass Equation
125
Mass interior to radius r is just the integral of the density in spherical coordinates Zr M r r04r02dr0
0
We usually prefer to write this as a di erential equation. Taking the deriva tive with respect to the upper limit of integration gives trivially
dM dr
4r2
Equation of State
If we could just nd an algebraic relation between pressure and density
P
we would be done 3 equations for the 3 unknowns P M as a function of the independent variable r. In real life however pressure depends not only on density but also on temperature and composition. For a mixture of a perfect gas and radiation we have 4 P Pgas Pradiation kT 1 3 aT remember our calculation of radiation pressure in 3.3.4 and of radiation energy density in 3.5.4 . Luckily it is often true that Pradiation Pgas so that the second term can be neglected and that we can either i derive from other physics or ii make a good guess about how T in the rst term varies with . Further in many cases of interest is constant. Then we will have arrived at a so called barytropic equation of state P P . Let us now make that assumption. Later we will catalog the actual cases for which it occurs. 126
Here n which need not be an integer is the so called polytropic index. The 1 can be thought of as arising from the perfect gas law weird notation 1 n
P T
1
T
or
Tn
But that is just notational history. It turns out that main sequence stars are pretty well modeled by n 3 polytropes. That is the run of temperature and density in the star roughly follows T 13 So the value n 3 is a good one to keep in mind as we proceed although we will meet other values later. Hydrostatic equilibrium dP GM dr r2 Mass dM 4r2 dr
1 d r2 dP r2 dr dr
127
4G
P
Change to dimensionless units
r where a some scale distance dimensionless length a dimensionless density function actually very temperature like
c n
where
P
Substitute for P
1 1 1 d a22 1 d K 1 n 1 n 4G c n c a22 a d c n a d 1 1 d 2 K 1 n d 1 n n d 4G c n c a22 d c n 0 1 2 1 d 2 d 4 Ga n( A 1 n 2 d d n 1 n c K We now can see that an inspired choice for the length scale a would be
a2
1 n c n K 4G
1 n
and we obtain the canonical form of the Lane Emden equation 1 d 2 d n 2 d d It is a common procedure in physics to convert equations to mathematical form like this
128
0 d
d
0 0
00 d
2
d
In practice there are better numerical methods than this but they reduce to this one conceptually. What are the boundary values 0 and 00 First 0 1 by de nition of c see above. Second we know that dP dr 0 at r 0 since GM r2 local acceleration of gravity goes to zero there. Thus from
P
we get 0 dP dr
1d 1 n 1 n dr
n 1 n d dr
0 0
n 1
must vanish.
129
With these boundary conditions you might enjoy deriving the rst terms in the power series expansion for namely 1 2 n 4 1 6 120 Numerical integration of the kind just described can give these so called Lane Emden functions to any desired accuracy. Here is a graph for n 0 1 5 3 0 and 3 5.
1 .8 .6 .4
n=0
Lane-Emden function
.2 0
n= 3.5 n= 3
1.5 n=
4 6 scaled radius
10
r a
2 nd
However here is a great trick if we multiply the Lane Emden equation by 2 and integrate from 0 to we get Z 2 d 2 n d d 0 So we can immediately read o
M r
4a3
c
d r a 3 2 n 1 K 1 p G 4
2 d
3 n 2n
2 d
r a
d
The total mass is obtained by setting 1 stellar surface. To get phys ical masses and radii we thus need tabulated numerical values for 1 and 2 d d you could in principle read these o the above graph of the 1 1 Lane Emden functions
n
0 05 10 15 20 25 30 35 40 45
1
2 4494 2 7528 3 14159 3 65375 4 35287 5 35528 6 89685 9 53581 14.97155 31.83646
2 1
d 1
4 8988 3 7871 3.14159 2.71406 2.41105 2.18720 2.01824 1.89056 1.79723 1.73780 131
1 0000 1 8361 3 28987 5 99071 11 40254 23 40646 54 1825 152 884 622 408 6189 47
You might wonder what happens if you integrate the Lane Emden equation beyond its rst root 1 Dont even think about it If is negative then the density is negative which is completely unphysical. The equation is perfectly happy to be truncated at 1 since 0 implies P 0 which is the correct surface boundary condition. Recapping the polytrope stellar model gives 2 algebraic relations among the 4 quantities K c M and R. Nowadays we might view M and K as the independent variables K which puts a scale on how hot the star is deriving from nuclear theory and use the model to determine c and R. Historically before nuclear processes were understood people used measured values for M and R and then derived c and K which were otherwise unknown. From either viewpoint once these 4 quantities are known we can go on to calculate how all quantities of interest vary with radius for example by using the graphs of the Lane Emden functions given above. Also one can easily derive relations like
center
3d 1 d
1
mean density of polytrope central pressure central temperature using the perfect gas law
132
2 0 From the numerical solution 1 6 90 1 1 2 02 and Thus M 31 g a3 c 7 9 10 2 0 41 1 and the scale length a 1 01 1010 cm
0 0424.
76 7 g cm
c
3 3 2
1 41 g cm
central pressure Pc 1 25 1017 dyne cm central temperature Tc 1 97 107 K We also get the constant K 3 85 1014 cgs in P
4 3.
133
The perfect gas law involves only the number density of particles N V not and is P N V kT
2 PV
Now the 1st Law of Thermodynamics which is really just conservation of energy says that when you squeeze a gas from volume V to volume V dV smaller volume the increase in its internal energy is just the work you have done squeezing it dE PdV The minus sign is because the internal energy increases with decreasing volume. Combining the last two equations taking a di erential of the rst 2 PdV
V dP
2 V dP dP P
PdV 1 2 PdV 2 dV V
Integrating we get
2 constant V 1
But since the density of a xed quantity of gas N particles varies inversely with its volume this is just 2 P 1 which is polytropic with index n 2. The most common case 3 5 3. Notice that as the number of degrees of freedom gives n 1 5 P 134
increases the polytropic index increases. In fact the limiting case of an isothermal gas P T T constant so P corresponds to 1. This is because one can view the work compressing the gas as being spread over an in nite number of internal degrees of freedom resulting in no increase in temperature.
constant
53
where all uid elements have the same constant. This is just what is needed for the validity of a polytropic model with n 3 2. The Sun is not fully convective most of it is stably strati ed with the deeper denser material being on a lower adiabat smaller value of constant in above equation. That is why the run of density and pressure in the Sun is better described by P 4 3 polytropic index n 3 than by P 5 3 135
polytropic index n 1 5 even though the material in the Sun is monatomic and fully ionized. The Sun has an outer convective envelope only in the last 1 6 or so of its radius. Low mass stars 0 3M are almost completely convective so they are good n 1 5 polytropes. Also stars of all masses go through an initial convective phase called the Hayashi phase before they settle down to the main sequence. This phase can typically last several million years. It too is well described by the Lane Emden equations for an n 1 5 polytrope.
temperature T entered only as an intermediate step in getting a relation between P and E V energy per volume namely
2E 3V
0E
perfect gas.
0
generalization. 1 PdV
0
PdV dE 1 1 0 dV V
Integrating gives
V dP
1 dP 0 P
0 1
constant V
0 1
P
with
0
PV
Now back to degenerate electrons. In our discussion of quantum phase space density in Section 3.5.2 we already saw that fermions because of the Pauli exclusion principle have a maximum phase space density of 1 per quantum unit h3 of phase space. Actually its 1 for each so called spin state 137
just as when for photons we counted each polarization separately. Electrons 1 and 1 . have two spin states called 2 2 Recall that phase space volume is the product of actual space volume and momentum space volume d3xd3p. Because of the limitation on phase space density if we force some electrons into a smaller volume d3x they must push out into a larger momentum volume d3p. The lowest energy state of an electron gas the state it will take on in practice at zero or low temperature is when the electrons ll all momentum states in a sphere out to some radius pF in momentum space called the Fermi momentum and none beyond. Thus for N electrons in a volume V we have the phase space density
4N 3 V 3 pF
p3 F
3 h3n 8 e
2 h3
Now to get the energy per volume E V we integrate over the density distri bution in momentum space putting in Ee p2 2me as the energy of each electron. Note that this is valid for non relativistic electrons E Z E N d3p Z pF p2 2 4p2dp 4 p5 F e 3 3 V 5 h me 0 2me h Now what about the relation between pressure P and energy density E V Actually degeneracy never comes into this. It is just the same cal culation we did in Section 3.3.4 but for a nonrelativistic particle moving at velocity V . Go back and look at the equations around these gures
138
L L
1 p2 3 ALme
2 E 3 V
or
2E 3V 5 3. So from the previous discussion of 0 we know right away that P Another way to verify this and also get the constant in the relationship is by substituting the above expression for E V in terms of p5 F and also substituting the earlier expression for p3 F in terms of ne . The result is the amazing universal formula 3 2 3 h2 1 3 P 20 m n5 non relativistic electrons. e e
Something di erent happens if the electron gas is so compressed that the Fermi momentum pF starts approaching mec i.e. the electron velocity starts approaching c. Then the above assumptions Ee p2 2me and ve p me start breaking down In the extreme relativistic limit the above calculation of P goes over to being exactly the same as we previously did for photons so
1E 3V
1 3
139
4 3
43
2c p4 h3 F
So the same two substitutions as before give 1 3 1 3 hc n4 3 P 8 relativistic electrons. e The crossover between these two limits can be roughly taken to be where they are equal 1 3 1 3 hc n4 3 1 3 2 3 h2 n5 3 e 8 20 me e 3 4 2 h ne 3 5 m c e The length h me c 2 42 10 10 cm is called the Compton wavelength of the electron. An electron forced into a box smaller than about 2 5 Compton wavelengths on a side we see becomes relativistic. This is of course a conse quence of the Heisenberg uncertainty principle relating positional localization to momentum de localization. A nal detail for our astrophysical applications is to relate ne the electron density to the total mass density. Except in the case of hydrogen which is not present in white dwarf stars anyway each electron is accompanied by exactly one proton and by very nearly one neutron. Thus
for hydrogen 2 for everything else. If a gas contains a mass fraction X of hydrogen X H then you can readily work out that
e 1 2 X
Elsewhere you might see not e which is de ned as the mean molec ular weight per particle not per electron. If Y He is the helium mass fraction and Z 1 X Y is the mass fraction of everything heavier than He then a good approximation is 2 1 3X
1Y 2
Here is a handy table of values for mixtures of H and He Pure H gas Hydrogenless gas Cosmic gas
e 1 e 2 X 0 74
1 2
1 21 Y 2 Y
0 26
e 1 15
0 60
Write these as
Pe Ke
with n 3 2 for the nonrelativistic case and n 3 for the very relativistic case. Then from polytrope theory Section 5.2.6 a white dwarf radius and mass are 1 2 n 1 K e 1 n 2n R 1 4G c 3 2 n 1 K e 3 n 2n 2 0 M 1 1 4 4G c
1
2 1
R M
So M R 3 with c 107 g cm 3 when M 1 5 M and the more mas sive a white dwarf is the smaller it is
1
6 897 142
2 1
R M
2 018
3 347 104 km c 106 g cm 3 1 457 M 2 e 2
13
e 2
23
Isnt it peculiar that M is independent of c It has a xed value 1 457M if e 2 that is called the Chandrasekhar mass MCh. What does this mean Recall that the extreme relativistic case is a limiting case as c 1. Thus the meaning is that as c 1 we get driven to the relativistic case with M MCh . That is the maximum mass that electron degeneracy pressure can possibly support. For higher masses there are simply no solutions Another way of understanding this is to look at the mass radius diagram plotting both the polytropes n 3 2 and n 3 and also a more complicated model that goes smoothly from the non relativistic to extreme relativistic limits
143
-1.5
Van Maanen 2 o 2 Eri B AC 70 8247
-2.0
log10(R/R . )
n = 3/2 polytrope
-2.5
-3.0
-3.5 -0.6
-0.4 -0.2
0.2 0.4
log10(M/M . )
0.6
So over this range the radius for a given mass decreases and at R 0 M MCh the relativistic electrons can no longer support the star. The points plotted on the curve are observational measurements of actual white dwarf stars and demonstrate that our theory is basically correct even accu rate quantitatively Note that e 2 is appropriate because X 0 the star must have burned all its hydrogen en route used up all nuclear fuel. In terms of fundamental constants we can write 3 2 c 1 MCh 3 10 h G m22
p e
144
where
1 2 h c mPlanck G This so called Planck mass has the numerical value 0.22 g. So MCh de pends only on mPlanck and mp even though it is electrons that support the star. Nowhere does me enter the formula. In the relativistic limit the rest mass does not enter the E PV relation. Thus me does not enter into the calculation of electron pressure.
This is of course fanciful since stars are always so hot that molecules are destroyed. But there is a point of principle to understand here. We can use thermodynamic results and the pressure equation to deduce a virial theorem relating to stellar structure. Start from the equation of hydrostatic equilibrium dP GM r r dr r2 4 r3 so that and de ne V r volume occupied by gas inside radius r 3 dV volume of dr shell containing mass r 4r2 dr. 145
V r dP dr dr
i.e.
V r dP
1 rZ R GM r dM r 3r 0 r
where U is the total gravitational potential energy. Hence integrating by parts 1 U PV r R rZ R P r dV r 0 3 r 0 At r 0 V 0. At r R P 0. The surface of a star is approximately a vacuum. Therefore R PV r 0 r 0 Hence
U 3
Zr
r
P dV
This is the most general form of the stellar structure virial theorem. For an ideal gas the energy per volume u is
We formerly wrote this as u E V when it was assumed constant over the volume V . Thus Z Z P dV 1 u dV 1 E 146
U 3
1E 0
5 and the total internal Note that for a perfect monatomic gas 3 energy unit mass is just the total kinetic energy of the gas particles i.e. E T . So U 3 1E 0 is equivalent to U 2T 0 the same as self gravitating particle virial theorem. Theorems are consistent fortunately The total energy of a star is
Etot E U
so
Etot E 3
We now see that if if 4 3 4 3
1E
3 3 3
4E 4 U 1
Etot 0 Etot 0
4 . Oth We see that stars are stable only if their adiabatic index exceeds 3 erwise they are unstable to converting their internal energy into expansion velocity they blow themselves apart
Actually this estimate is rather inaccurate 1 because at this density nuclear forces strong force between neutrons are appreciable and this aids gravity thus decreasing the limiting mass and 2 because the Newtonian gravitational potential surface c2 1 so Newtonian gravity is a poor approx imation and we should use General Relativity. The physics overall is rather uncertain but the limiting mass of a neutron star is probably 3 M . Un til we know the equation of state of a neutron uid this mass will remain uncertain. What is the radius of a neutron star Neutron stars are denser than white dwarf stars. In the limit the central density should correspond to the neutrons almost touching. The neutron neutron separation is then 1 fm 10 13 cm so the neutron number density 1039 cm 3 and 2 1015 g cm 3. Hence if M 1 M R 10 km. We know that neutron stars exist because of evidence of pulsars. We know that white dwarfs exist too because we can see them optically e.g. Sirius B. Pulsars are distinguished by the regular arrival of radio pulses separated by intervals of a few ms 0.5 ms shortest to a few seconds 5 sec longest.
4 Time (s)
Chart record of individual pulses from one of the rst pulsars discovered PSF 0329 54. They were recorded at a frequency of 410 MHz and with an instrumental time constant of 20 ms. The pulses occur at regular intervals of about 0.714 s.
149
The period is quite regular but slowly decreasing due to spindown of the neutron star. The ux and pulse shape are somewhat variable but the long time averages are stable. Some pulsars put out optical and x ray pulses as well as radio pulses. What makes pulsars pulse The answer is thought to be that the pulses are from electrons trapped in the magnetic eld of the rotating neutron star. The idea is that the pulses are thus emitted in a cone rotating with the neutron star. Suppose the period is P 2 . Then we can derive a limit to the neutron star radius r. At the equator the centrifugal force 2r per unit mass had better not exceed the gravitational force GM r2 per unit mass.
to
+
rotation axis
neutron star
charge
Therefore
GM r2 for stability. So for a given P we require P 2 3 2 1 3 GMP 1500 km r 42 sec 1 white dwarf radius Crab pulsar P 33 ms r 150 km 10
2r
1 white dwarf radius Pulsar 1937 21 P 2 ms r 24 km 100 These objects cant be white dwarf stars periods too short and they cant be planets or the rate of loss of energy would slow them too quickly so we infer that they are neutron stars.
150
We see neutron star slow down at a rate that is consistent with energy loss by electromagnetic radiation due to rotating magnetic dipole. This requires very strong magnetic elds 1012 Gauss believed to be present.
151